Medulloblastoma-associated DDX3 variant selectively alters the translational response to stress

Abstract

DDX3X encodes a DEAD-box family RNA helicase (DDX3) commonly mutated in medulloblastoma, a highly aggressive cerebellar tumor affecting both children and adults. Despite being implicated in several facets of RNA metabolism, the nature and scope of DDX3's interactions with RNA remain unclear. Here, we show DDX3 collaborates extensively with the translation initiation machinery through direct binding to 5'UTRs of nearly all coding RNAs, specific sites on the 18S rRNA, and multiple components of the translation initiation complex. Impairment of translation initiation is also evident in primary medulloblastomas harboring mutations in DDX3X, further highlighting DDX3's role in this process. Arsenite-induced stress shifts DDX3 binding from the 5'UTR into the coding region of mRNAs concomitant with a general reduction of translation, and both the shift of DDX3 on mRNA and decreased translation are blunted by expression of a catalytically-impaired, medulloblastoma-associated DDX3R534H variant. Furthermore, despite the global repression of translation induced by arsenite, translation is preserved on select genes involved in chromatin organization in DDX3R534H-expressing cells. Thus, DDX3 interacts extensively with RNA and ribosomal machinery to help remodel the translation landscape in response to stress, while cancer-related DDX3 variants adapt this response to selectively preserve translation.

Keywords: CLIP-seq; DDX3; DDX3X; RNA helicase; medulloblastoma.

Figure 1. DDX3 binds the 5′ leader of mRNA and specifically interacts with the 18S rRNA

(A) Autoradiogram of ³²P labeled RNA UV-crosslinked to FLAG-DDX3 (top) and Western blot (WB) analysis of DDX3 enrichment (bottom). β-Actin, loading control. (B) Distribution of RT stops identified in the DDX3 and DDX3^R534H iCLIP experiments. Percentage of unique RNA crosslinked in each RNA class is noted. (C) Distribution of RT stops mapped to indicated mRNA regions. (D) Metagene analysis of normalized iCLIP RT stop density from DDX3 (green) or DDX3^R534H (red) experiments. Scaled 5′UTR regions and 50 nts up and downstream of the 5′UTR are plotted. TIS, translation initiation site. CDS, coding sequence. (E) Normalized iCLIP RT stop density from DDX3 and DDX3^R534H iCLIP mapping to the full rRNA transcribed region. (F) Zoom of e at the most well bound site on the 18S rRNA with the solvent accessibility of the 18S rRNA annotated below at single nucleotide resolution (calculated using PyMOL). (D–E) iCLIP RT stops were normalized for library size and calculated as counts in 10 million. (G) Structure of the human 40S ribosome with the three binding sites of DDX3 in f highlighted in green and indicated with arrows.

Figure 2. DDX3X mutations in SHH subtype MB exhibit impaired translation

(A) Conserved regions in the DDX3 family are shown in dark gray with canonical DEAD-box protein domains in light gray. Medulloblastoma-associated mutations in either the SHH (top) or WNT (bottom) subtype are indicated by orange arrows with recurrent mutations colored red. (B and C) Unsupervised consensus clustering (k = 2) of RNA-seq-derived gene expression data (FPKM) from DDX3X-mutated and DDX3X wild type (WT) SHH-subtype medulloblastoma cases accrued from Toronto (A) or the German Cancer Research Center (DKFZ) (C). (D and E) Selected enrichment plots identifying genes associated with translation initiation and ribosome assembly enriched in wild-type relative to DDX3X-mutated medulloblastoma cases in both Toronto (C) and DKFZ (D) cohorts. GO x = gene ontology (MsigDB c5), CP = canonical pathways (MSigDB c2).

Figure 3. DDX3 catalytic activity is required for proper translation response to stress

(A) (Top) Metagene analysis of normalized iCLIP RT stop density from DDX3 (green) and DDX3^R534H (red) iCLIP experiments with (solid lines) or without (dotted lines) arsenite treatment. 5′UTR, CDS, and 3′UTR regions were each scaled to 200 bins. iCLIP RT stops were normalized for library size and calculated as counts in 10 million and the average RT stop density is shown (Bottom) Metagene plot of iCLIP RT stop density of arsenite-treated cells relative to untreated cells from DDX3 (green) and DDX3^R534H (red) iCLIP experiments. (B) (Top) Normalized ribosome footprint density along a metagene, averaged from entire translating mRNAs, scaled to 200 bins each for 5′UTR, CDS, and 3′UTR, normalized for library size, and calculated as counts in million mapped reads. Solid lines, arsenite-treated cells. Dotted lines, arsenite-untreated cells. (Bottom) Ribosome footprint density plot showing ribosome engagement along mRNA in arsenite-treated cells expressing Rluc (blue), DDX3 (green), DDX3^R534H (red) relative to untreated cells. (C) Average ribosome footprint density (normalized for library size and shown as read count per million mapped reads) around iCLIP RT stops in CDS excluding 200 nt after the start codon. Shaded area = mean ± standard error.

Figure 4. Gene-specific translation response to stress

(A) Frequency distribution of translation efficiency (TE) of genes from cells expressing Rluc (black), DDX3 (green) or DDX3^R534H (red) with sodium arsenite treatment relative to no treatment. Median TE fold reduction of Rluc, DDX3, or DDX3^R534H –expressing cells are shown. Z score was calculated from TE of Rluc-expressing cells: grey area: Z-score > 1.5 or < -1.5 and bracketed white area: Z-score between -1.5 and 1.5. (B) iCLIP RT stops in arsenite-treated cells relative to those in cells without arsenite treatment. Genes were grouped by Z-scores as calculated in a. (C) GC content (%) in genes grouped by Z-score in a. Genes were grouped by Z-scores as calculated in a. Dashed line at zero equals to no change in iCLIP RT stops induced by arsenite. Areas above and below the zero dashed line represent iCLIP gain and loss induced by arsenite, respectively. Each gene group shows similar range of variance; Standard deviations for gene groups in DDX3-expressing cells are 1.396 (Z < −1.5), 1.495 (−1.5 < z < 1.5), and 1.891 (z > 1.5). Standard deviations for gene groups in DDX3^R534H-expressing cells are 1.229 (Z < −1.5), 1.408 (−1.5 < z < 1.5), and 1.756 (z > 1.5). Error bars equal to mean ± 95% CI. ****P < 0.0001, ns (not significant) P > 0.05 calculated by ANOVA. (D) GO BP terms enriched in Z < −1.5 (left) and Z > 1.5 (right) gene sets as grouped in a. Dashed line, Benjamini-corrected P-value = 0.05. (E) A model for the consequence of loss of DDX3 catalytic activity in translation response. Translation is globally repressed concomitant with stress granule formation in response to stress such as arsenite, which is facilitated by wild-type DDX3 expression (middle). When the catalytically impaired DDX3^R534H is expressed, select transcripts escape the global translational repression caused by stress (right).